Anthropic adds reporting tools to show where Claude spending goes

Anthropic has introduced a new set of reporting tools for business customers that want a clearer view of what their AI budgets are producing, according to TechCrunch. The move matters because companies are increasing spending on agentic coding and AI assistants while finance and operations teams are asking harder questions about return on investment.

The tools are aimed at making token usage less opaque. Instead of only seeing that teams are spending on Anthropic products, managers can examine what work was done, what it cost and where the process ran into problems.

What Anthropic’s new reports show

Anthropic’s reporting suite uses internal analytics and employee surveys to help managers review a team’s work across a selected date range, TechCrunch reported. The reports can show what the team produced, how much that work cost and where workflows broke down.

The company also says the reports will provide recommendations for improving results. Anthropic says it will not use the underlying data for model training by default.

For companies already using Claude, the practical value is straightforward: AI usage becomes easier to discuss in the same language as other operating costs. A manager can look beyond adoption and start asking whether a workflow is saving time, creating review bottlenecks or simply moving costs from one budget line to another.

Anthropic is also creating pooled group budgets, which are designed to help teams manage token spending more directly. That matters for organizations where AI usage may start in engineering or product teams but quickly becomes a shared finance issue.

Why AI spend is becoming a management problem

The timing reflects a broader shift in enterprise AI. TechCrunch reported that Anthropic has been focused on keeping enterprise users engaged, and the company’s investor materials pointed to strong business momentum around Claude Code. According to the report, Anthropic recorded an operating profit in the second quarter, while spending from the same customers rose sharply from Q3 2025 to Q3 2026.

But the article also notes a tension: many businesses are still trying to work out how to deploy AI effectively. Usage among venture-backed companies is helping drive token spending, while most Fortune 500 companies are not seeing major gains from AI rollouts, according to TechCrunch.

That gap is where reporting becomes important. For creators, agencies, freelance operators and small businesses, the lesson is not that every team needs enterprise dashboards. It is that AI costs should be tied to a workflow, an owner and a measurable output before spending scales.

A simple version of the same discipline can work without enterprise tooling:

• Decide which task the AI tool is meant to improve.
• Track usage by project or team, not just by individual enthusiasm.
• Review where human editing, approvals or handoffs still slow the work down.
• Set a budget before usage becomes routine.

Anthropic’s new tools formalize that process for larger customers, but the management problem is familiar to smaller teams too.

Claude Code’s enterprise position is part of the story

The reporting launch is closely tied to the growth of agentic coding tools. Jellyfish, a company that builds tools for software developers, released a report based on usage by 276,000 of its customers. According to TechCrunch, tools built on Anthropic models, including Claude Code, were the most widely used in that survey and accounted for nearly a third of proposed code changes submitted for review.

Jellyfish also found that spending on those tools had grown tenfold since the start of the year, with the cost now equal to about 2% of total engineering cost, TechCrunch reported. Nicholas Arcolano, Jellyfish’s chief scientist, told the outlet that even if 2% remains small, that rate of growth is something many finance chiefs had not planned for.

That puts Anthropic’s reporting tools in a competitive context. OpenAI’s Astra model was described by TechCrunch as a potential threat after its release this summer, but Jellyfish’s data has not shown it slowing Anthropic’s adoption numbers so far. For more on the enterprise battle around model customers, see our earlier coverage of OpenAI discounts aimed at Anthropic’s enterprise accounts.

From hands-on help to self-serve oversight

TechCrunch frames the new reports as a lighter version of the forward-deployed engineering teams that major AI labs use to help customers adopt their products. Those teams embed AI consultants inside customer organizations to get deployments working.

Anthropic’s reports could give customers some of that guidance in product form: usage visibility, workflow diagnosis and recommendations without requiring the same level of hands-on support.

For buyers, the key question is whether these reports can identify where AI tools are genuinely becoming part of repeatable work. Jellyfish’s Nikolas Albarran told TechCrunch that switching costs rise once companies standardize workflows around a particular system, because each tool has its own quirks.

That is the strategic layer behind a seemingly practical feature. If Anthropic can help customers understand and improve their Claude workflows, it may also make those workflows harder to replace.